accelerator

Analyses a focused set of screens in a running web application via the Playwright MCP server. Captures detailed state, screenshots, and computed values. Call browser-analyser when you need to extract HOW a screen behaves, not to enumerate WHERE things are.

Locates routes, screens, and DOM-level component presence in a running web application via the Playwright MCP server. Call browser-locator when you need to enumerate WHERE things appear in the rendered UI, not to extract their detail.

Analyses codebase implementation details. Call the codebase-analyser agent when you need to find detailed information about specific components. As always, the more detailed your request prompt, the better! :)

Locates files, directories, and components relevant to a feature or task. Call `codebase-locator` with human language prompt describing what you're looking for. Basically a "Super Grep/Glob/LS tool" — Use it if you find yourself desiring to use one of these tools more than once.

codebase-pattern-finder is a useful subagent_type for finding similar implementations, usage examples, or existing patterns that can be modeled after. It will give you concrete code examples based on what you're looking for! It's sorta like codebase-locator, but it will not only tell you the location of files, it will also give you code details!

The research equivalent of documents-analyser. Use this subagent_type when wanting to deep dive on a research topic. Not commonly needed otherwise.

Discovers relevant documents in meta/ directory (We use this for all sorts of metadata storage!). This is really only relevant/needed when you're in a reseaching mood and need to figure out if we have random thoughts written down that are relevant to your current research task. Based on the name, I imagine you can guess this is the `documents` equivalent of `codebase-locator`

Generic review agent that evaluates code or plans through a specific quality lens. Spawned by review orchestrators with lens-specific instructions and output format injected at spawn time.

Do you find yourself desiring information that you don't quite feel well-trained (confident) on? Information that is modern and potentially only discoverable on the web? Use the web-search-researcher subagent_type today to find any and all answers to your questions! It will research deeply to figure out and attempt to answer your questions! If you aren't immediately satisfied you can get your money back! (Not really - but you can re-run web-search-researcher with an altered prompt in the event you're not satisfied the first time)

View, create, or edit Accelerator plugin configuration. Manage document templates.

Prepare a repository with the directories and gitignore entries that Accelerator skills expect. Safe to run repeatedly.

Apply pending Accelerator meta-directory migrations to bring a repo into line with the latest plugin schema. Destructive by default but guarded — refuses to run on a dirty working tree and prints a one-line preview per pending migration before applying.

Interactively create an architecture decision record (ADR). Use when the user wants to document an architectural decision, technology choice, or significant design decision. Guides through context gathering, options analysis, and consequence documentation.

Extract architecture decision records from existing meta documents (research, plans). Scans documents for implicit or explicit architectural decisions and converts selected ones into formal ADRs. Use when decisions are buried in research or planning documents and need to be captured formally.

Review an architecture decision record for quality and completeness, then accept, reject, or suggest revisions. Enforces ADR immutability — only proposed ADRs can be modified, accepted ADRs can only transition to superseded or deprecated. Use when a proposed ADR is ready for review, or when an accepted ADR needs to be deprecated.

Compare two design inventories produced by inventory-design and emit a structured gap artifact whose prose paragraphs satisfy the extract-work-items cue-phrase contract. Use after running inventory-design for both a current and target design surface. The resulting gap artifact under meta/design-gaps/ feeds directly into /accelerator:extract-work-items to produce actionable work items.

Generate a structured design inventory for a frontend source — tokens, components, screens, and features — by crawling it with code analysis, live Playwright inspection, or both. Use when you need to capture a snapshot of a current or target design surface before running analyse-design-gaps. Produces a dated artifact directory with an inventory.md and screenshots/. Re-running for the same source-id supersedes the prior snapshot without losing it.

Generate a comprehensive pull request description following the repository's standard template. Use when the user wants to create or update a PR description.

Respond to pull request review feedback interactively, working through each item with verification and code changes. Use when the user wants to address PR review comments.

Review a pull request through multiple quality lenses and present a compiled analysis with inline comments. Use when the user wants a thorough PR review.

Create detailed implementation plans through interactive, iterative collaboration. Use when the user needs to plan a feature, refactoring, or task.

Execute an approved implementation plan from the configured plans directory. Use when the user wants to implement a plan phase by phase with verification.

Review an implementation plan through multiple quality lenses and collaboratively iterate based on findings. Use when the user wants to evaluate a plan before implementation.

Interactively stress-test an implementation plan by grilling the user on decisions, edge cases, and assumptions to find issues, inconsistencies, and gaps before implementation begins.

Validate that an implementation plan was correctly executed by verifying success criteria and identifying deviations. Use after implementing a plan to verify correctness.

Conduct comprehensive codebase research by spawning parallel sub-agents and synthesising findings into a research document. Use when the user needs to deeply understand a codebase area or answer technical questions.

Create VCS commits for session changes. Use when the user wants to commit their work with well-structured, atomic commits.

Open the accelerator meta visualiser. Launches the companion-window server in the background and returns a URL. Subcommands stop and status manage the running server.

Interactively create a well-formed work item. Use when capturing a feature, bug, task, spike, or epic as a structured work item in meta/work/.

Extract work items in batch from existing documents (specs, PRDs, research, plans, meeting notes, design docs). Use whenever the user wants to capture, pull out, or convert requirements, work items, user stories, bug reports, or actionable tasks from existing files into structured work items in meta/work/ — even if they don't say "extract" explicitly.

List and filter work items from the configured work directory. Use when discovering what work items exist, filtering by status/type/priority/parent/tag, or viewing the work item hierarchy.

Interactively refine a work item by decomposing it into children, enriching it with codebase context, sharpening its acceptance criteria, sizing it, or linking it to dependencies. Use after a work item has been drafted and before planning begins.

Review a work item through multiple quality lenses and collaboratively iterate based on findings. Use when the user wants to evaluate a work item before implementation or escalation.

Interactively stress-test a work item by grilling the user on scope, assumptions, acceptance criteria, edge cases, and dependencies to surface issues, gaps, and flawed assumptions before implementation is planned.

Update fields (status, priority, tags, parent, etc.) of an existing work item. Use to transition status, change priority, manage tags, or edit any frontmatter field. No transition enforcement — arbitrary changes are allowed.

Use this skill only when the user explicitly invokes /attach-jira-issue to upload one or more local files as attachments to a Jira issue. This is a write skill with irreversible side effects — it must never be auto-invoked from conversational context. Shows a preview of what will be uploaded and requires explicit confirmation before POSTing.

Use this skill only when the user explicitly invokes /comment-jira-issue to add, list, edit, or delete comments on a Jira issue. This is a write skill with irreversible side effects — it must never be auto-invoked from conversational context. Subcommands: add (post a new comment), list (fetch all comments with pagination), edit (update an existing comment), delete (remove a comment — irreversible). Write subcommands show a payload preview and require explicit confirmation before calling the API.

Use this skill only when the user explicitly invokes /create-jira-issue to create a new Jira issue. This is a write skill with irreversible side effects — it must never be auto-invoked from conversational context. Accepts a project key, issue type, summary, optional Markdown body, and optional fields (assignee, priority, labels, components, parent, custom fields). Converts the body to ADF, shows a payload preview, requires explicit confirmation, then POSTs to Jira and returns the new issue key.

Set up the Jira Cloud integration for this project. Verifies credentials against a real Jira Cloud tenant, discovers the tenant's custom-field catalogue and project list, and persists the results under `<paths.integrations>/jira/` (default `meta/integrations/jira/`) as team-shared, version-controlled JSON caches. Idempotent: safe to re-run after credential or project changes.

Use this skill whenever the user wants to search, list, or filter Jira tickets — by assignee, status, label, project, type, component, reporter, parent, or free text — even if they say 'find', 'show me', 'what's open', 'list my tickets', or similar phrasing rather than 'search Jira'. Composes safe JQL from structured flags, executes a paginated search against a Jira Cloud tenant, and renders a summary table of the results. Supports --render-adf to convert ADF descriptions to Markdown inline. Prefer this skill over raw JQL whenever the user's intent maps to a structured flag.

Use this skill when the user asks about a specific Jira issue by key (e.g. PROJ-123, ENG-456) — for viewing the description, status, comments, transitions, or any other field. Trigger when the user says 'look up', 'check on', 'tell me about', 'what's on', or 'what is the status of' a key, or asks any direct question about an issue they reference. Do NOT trigger when an issue key appears incidentally inside other prose (commit messages, code review comments, release notes), where the user is talking about the issue rather than asking to fetch it.

Use this skill only when the user explicitly invokes /transition-jira-issue to move a Jira issue through its workflow by state name. This is a write skill with irreversible side effects — it must never be auto-invoked from conversational context. Accepts an issue key and target state name (case-insensitive). Shows a transition preview and requires explicit confirmation before posting.

Use this skill only when the user explicitly invokes /update-jira-issue to modify an existing Jira issue. This is a write skill with irreversible side effects — it must never be auto-invoked from conversational context. Accepts an issue key and at least one mutating flag (summary, body, priority, assignee, labels, components, parent, custom fields). Shows a payload preview with explicit set-vs-update label semantics, requires explicit confirmation, then PUTs to Jira.

Architecture review lens for evaluating structural integrity, coupling, cohesion, and evolutionary fitness. Used by review orchestrators — not invoked directly.

Work-item review lens for evaluating unambiguous communication — referent clarity, internal consistency, jargon handling, and actor/outcome identification. Used by review orchestrators — not invoked directly.

Code quality review lens for evaluating design principles, error handling, complexity, testability, and maintainability. Used by review orchestrators — not invoked directly.

Compatibility review lens for evaluating API contract stability, cross-platform support, protocol compliance, and dependency management. Used by review orchestrators — not invoked directly.

Work-item review lens for evaluating structural and informational completeness — section presence, content density, type-appropriate content, and frontmatter integrity. Used by review orchestrators — not invoked directly.

Correctness review lens for evaluating logical validity, boundary conditions, invariant preservation, concurrency correctness, and state management. Used by review orchestrators — not invoked directly.

Database review lens for evaluating migration safety, schema design, query correctness, and data integrity. Used by review orchestrators — not invoked directly.

Work-item review lens for evaluating explicit capture of blockers, consumers, external systems, and ordering. Used by review orchestrators — not invoked directly.

Documentation review lens for evaluating documentation completeness, accuracy, and audience-appropriateness. Used by review orchestrators — not invoked directly.

Performance review lens for evaluating algorithmic efficiency, resource usage, and concurrency efficiency. Used by review orchestrators — not invoked directly.

Portability review lens for evaluating environment independence, deployment flexibility, and vendor lock-in avoidance. Used by review orchestrators — not invoked directly.

Safety review lens for evaluating data loss prevention, operational safety, and protective mechanisms against accidental harm. Used by review orchestrators — not invoked directly.

Work-item review lens for evaluating sizing, decomposition, and orthogonality of requirements. Used by review orchestrators — not invoked directly.

Security review lens for evaluating threats, vulnerabilities, and missing protections. Used by review orchestrators — not invoked directly.

Standards compliance review lens for evaluating project conventions, API standards, and accessibility. Used by review orchestrators — not invoked directly.

Test coverage review lens for evaluating testing strategy adequacy, test quality, and test architecture. Used by review orchestrators — not invoked directly.

Work-item review lens for evaluating whether Acceptance Criteria and requirements admit a concrete verification strategy — each criterion must be specific, measurable, and verifiable. Used by review orchestrators — not invoked directly.

Usability review lens for evaluating developer experience, API ergonomics, configuration complexity, and onboarding. Used by review orchestrators — not invoked directly.

Output format specification for plan review agents. Defines the JSON schema, field reference, severity emoji prefixes, and finding body format for plan reviews.

Output format specification for PR review agents. Defines the JSON schema, field reference, severity emoji prefixes, and comment body format for PR reviews.

Output format specification for work-item review agents. Defines the JSON schema, field reference, severity emoji prefixes, and finding body format for work-item reviews. Used by review orchestrators — not invoked directly.

4 hooks across 2 events